Nexus Eclipse

Resources

QA templates and guides you can use this week

Templates for the artefacts we actually write on engagements, plus reference guides for evaluation, automation shape, process, and cost. No email gate.

Templates

Guides

LLM as a Judge

What LLM-as-a-judge means in product QA: when a model can score another model’s output, when it shares the same failure modes, and how to keep a human in the loop.

LLM Evaluation Metrics

LLM evaluation metrics that belong on a product: task pass rates, format checks, groundedness, refusals, and version diffs — not a generic BLEU score or a public leaderboard.

Playwright vs Cypress

Playwright vs Cypress for product teams: how to choose a browser automation tool based on who will maintain the suite — not a religion, a partnership, or a feature checklist copied from a vendor site.

Shift Left Testing

Shift-left testing without the slogan: move checks earlier in the lifecycle, keep QA in the loop, and avoid using “shift left” as a reason to skip a test stage.

Test Pyramid

The test pyramid without the poster: many fast unit and API checks, fewer UI journeys, and a deliberate place for exploratory and AI evaluation — not an ice-cream cone of end-to-end tests.

QA Process

A practical QA process for product teams: intake, risk, cases, execution, defects, and release evidence — without a maturity-model poster or a fake certification.

QA Metrics

QA metrics that change a decision: coverage you can defend, escaped defects, flake rate, and residual risk — not a vanity dashboard or a fake quality score.

Smoke Testing vs Sanity Testing

Smoke testing vs sanity testing without the syllabus fight: a build-level ‘does it start’ check versus a focused check after a change — and how both sit in a real QA process.

Severity vs Priority

Severity vs priority in defect reports: how bad the failure is versus how soon you will fix it — with examples, and without a fake scoring model.

QA Outsourcing Cost

What actually drives QA outsourcing cost: scope, surfaces, duration, evidence bar, and access — not a public day-rate or a ‘cheap offshore’ SKU.

AI Testing Tools

How to choose AI testing tools for a product that uses a model: eval harness, tracing, judges, and red-team aids — a buying guide, not a numbered vendor ranking.

How to Choose a Software Testing Company

A buying guide for software testing companies: what to ask, what to ignore, and how Nexus Eclipse actually works — not a numbered ranking we invented.

How to Choose a Software Testing Company in the UK

A UK buying guide for software testing companies: hours, accessibility context, and what Nexus Eclipse will not pretend — not a ranked list of UK firms.

What these resources are for

Most teams do not fail for lack of a tool. They fail because nobody can point to what was supposed to be tested, what was tested, and what a tester should write down when something breaks. The five templates close that gap.

The guides are for the arguments that keep showing up: how to score a model, which browser runner to keep, how a QA process should run, what to measure, smoke vs sanity, severity vs priority, and what drives outsourcing cost. They are not a maturity model and not a ranking of other companies.

Templates are the same artefacts we produce onmanual and functional testingandQA consultingengagements — published so you can use them without hiring us first.

How to pick a starting document

If requirements exist and you cannot see which tests cover them, start with therequirements traceability matrix. If a release is coming and scope is fuzzy, use thetest plan template. If testers are executing from memory, use thetest case template. If bugs bounce because they are incomplete, use thebug report template. If the team cannot agree how testing should work at all, use thetest strategy document.

If the argument is “which runner” or “how many UI tests,” start withPlaywright vs Cypress or thetest pyramid. If the product has a model, start withLLM as a judge andLLM evaluation metrics.

A QA audit is the right next step when you already have documents and still cannot tell what a release actually covered.

FAQs

Are these templates free to use?

Yes. Copy them, download them, and adapt them to your tracker. There is no email gate. They are working documents, not a product.

Do I need all five documents?

No. A requirements traceability matrix and test cases cover most product teams. Add a test plan when a release needs a written scope. Add a test strategy when the approach itself is unclear. Use the bug report template whenever defects are landing incomplete.

What are the guides for?

The guides explain LLM evaluation, Playwright vs Cypress, shift-left, the test pyramid, QA process and metrics, smoke vs sanity, severity vs priority, outsourcing cost, AI testing tools, and how to choose a software testing company (including a UK version). They are not templates and not a numbered ranking of other firms.

Will you fill these in for us?

Yes, as part of a testing engagement. The templates are here so you can start without us. If you want Nexus Eclipse to write the plan, cases, or matrix on a live product, that is QA work — see Services.

Are these the same as Insights?

No. Insights are dated articles. Resources are reusable templates plus reference guides you can keep open while you work.

Talk to our QA team

Need these filled in on a real product?

Send the product URL and the next release date. We will say whether you need a matrix, a plan, or testers executing cases.